GPT-5.6 Terra (Non-reasoning): benchmark results
GPT-5.6 Terra evaluated with reasoning disabled. Provider: OpenAI. Released 2026-07-09. Access: API.
Aggregate Rating 1674 ± 13, rank #441 of 2168 rated models, from 116 benchmark results.
Strongest benchmark results
| Benchmark | Score | Metric | Percentile |
|---|---|---|---|
| OntoAct Reference - Mars (ReAct) | 13.91 | Mars Score (%) | 100 |
| K-Bench - Risk (Therapeutic v0) | 94.25 | Risk Score (%, Therapeutic v0 prompt) | 93.5 |
| LLM Milgram Obedience - Baseline Mean Breakoff Voltage | 78 | Mean last shock administered in valid baseline sessions (vol | 89.5 |
| UGI - Natural Intelligence | 52.63 | NatInt Score | 89.2 |
| K-Bench - Risk | 94.35 | Risk Score (%, Default v1 prompt) | 88.5 |
| LLM Milgram Obedience - Baseline Full-Obedience Rate | 0 | Valid baseline sessions ending at 450 V (%; lower is less ob | 86.8 |
| UGI - Writing | 48.2 | Writing Score | 85.8 |
| ParseBench - Charts | 63.75 | ChartDataPointMatch Score | 84.8 |
| UGI Leaderboard | 47.66 | UGI Score | 84.3 |
| BrainBench (EEG) - BrainAgent - Neurocognitive Assessment | 72.16 | EEG analysis score (0-100) | 83.3 |
| Chehre | 42.7 | Top-1 accuracy (%; share of the 2,111 videos where the model | 80 |
| Chehre - Recall@2 | 59 | Recall@2 (%; share of the two highest-rated human labels rec | 80 |
Interactive version: theaggregate.ai/model?slug=gpt-5-6-terra-non-reasoning · How It Works · Data refreshed daily, snapshot 2026-10-20.